ISSUE: Researchers and research sponsors alike have an interest in how expanding EHR data can be used to generate registry-like patient insights quicker and at lower cost with greater generalizability. Can EHRs replace clinical registries? Sharon Hensley of IBM Explorys will address the potential to replicate registry data efficiently through EHRs. Tracy Wang of the Duke Clinical Research Institute will argue for the continuing value of registries to answer research questions. Marc Berger, who headed real world data and analytics at Pfizer, will share the perspective of pharmaceutical companies that commission clinical research on the cost-benefit of each approach. The panelists will debate before the discussion is opened to audience participation.
OVERVIEW: Clinical registries, a well-established mechanism for collecting standardized data on a well-defined study population, are considered the gold standard for understanding natural history, quality of care, clinical effectiveness and safety. But registries require a dedicated research infrastructure, which can be expensive, time consuming and lack generalizability. Utilizing EHR data for research is appealing as sample sizes are large, data are readily available and populations are representative. However, limitations include selection bias toward sicker patients, issues with accuracy and completeness of data, privacy considerations, challenges identifying when patients are present in the health system, and misclassification errors due to dependence on coding.